feat: add BMAD Guide skill for automatic Claude Code installation

Integrates the bmad-guide skill into BMAD installation process to help
Claude stay on track with BMAD methodology.

Features:
- Phase navigation and workflow selection guidance
- Project level detection (0-4) for routing to correct workflows
- Common mistakes prevention (skipping phases, manual coding)
- Emergency recovery procedures when off track
- Quick reference tables and troubleshooting

Implementation:
- Created resources/skills/ directory with bmad-guide.md (428 lines)
- Added installBmadGuideSkill() to ClaudeCodeSetup class
- Skill auto-installs to ~/.claude/skills/ during setup
- Available globally across all BMAD projects
- Updated installation documentation

Usage:
- Automatic: Skill is installed when running npx bmad-method install
- Manual access: /bmad-guide command in any project
- Optional: Add BMAD Method config to ~/.claude/CLAUDE.md for automatic invocation

Benefits:
- Prevents Claude from jumping straight to coding without workflows
- Ensures proper phase progression
- Guides workflow selection based on project level
- Self-corrects when going off track
This commit is contained in:
Jonah Schulte 2025-12-27 22:06:37 -05:00
parent ee68b027fc
commit 747ba512aa
4 changed files with 616 additions and 0 deletions

View File

@ -50,6 +50,12 @@ your-project/
└── ... your code # maybe nothing else yet if a fresh new folder
```
**For Claude Code users**, the installer also sets up:
- **BMAD Guide skill** (`~/.claude/skills/bmad-guide.md`) - Helps Claude stay on track with BMAD methodology
- Access with `/bmad-guide` command
- Provides phase navigation, workflow selection, and process guidance
- Automatically prevents common mistakes like skipping phases or manual implementation
## Next Steps
1. **Read the [Quick Start Guide](../modules/bmm/quick-start.md)** to build your first feature

146
resources/skills/README.md Normal file
View File

@ -0,0 +1,146 @@
# BMAD Guide Skill
This directory contains the **BMAD Guide Skill** for Claude Code - a comprehensive reference that helps Claude stay on track with BMAD methodology.
## What's Included
- **bmad-guide.md** - The skill file that provides process navigation and workflow selection guidance
## Automatic Installation
When you run `npx bmad-method install` and select Claude Code as your IDE, this skill is **automatically installed** to `~/.claude/skills/bmad-guide.md`.
This means Claude will have access to the `/bmad-guide` skill in any project, helping it:
- Navigate BMAD phases correctly
- Choose the right workflow for each task
- Avoid common mistakes (jumping to coding, skipping phases, etc.)
- Follow proper story lifecycle
- Self-correct when going off track
## What the Skill Does
The bmad-guide skill acts as Claude's "GPS" for BMAD methodology:
### 📍 Phase Navigation
Quick reference for identifying current phase and what workflows are available
### 🎯 Project Level Detection
Helps determine project complexity (Level 0-4) to route to correct planning track
### 🔍 Workflow Decision Tree
Visual guide for choosing which workflow to use for any given task
### ⚠️ Common Mistakes Prevention
Clear DO/DON'T lists to avoid derailment from BMAD process
### 📚 Quick Reference
"I need to..." → workflow mapping table for fast lookup
### 💡 Troubleshooting
Solutions for common issues like "I'm not sure which phase I'm in"
### 🚨 Emergency Recovery
Course correction steps when Claude has gone off track
## How It Works
### Automatic Invocation
When combined with the proper CLAUDE.md configuration (see below), Claude will **automatically invoke** this skill:
- Before starting any task in BMAD projects
- When uncertain about which workflow to use
- Before implementing features manually
- When switching phases
- When detecting red flags (coding without workflows, etc.)
### Manual Invocation
You can also invoke it manually anytime:
```bash
/bmad-guide
```
## CLAUDE.md Configuration (Optional but Recommended)
For maximum effectiveness, add this to your `~/.claude/CLAUDE.md`:
```markdown
# BMAD Method (MANDATORY)
## Automatic Skill Invocation - THIS IS CRITICAL
**IMMEDIATELY invoke `/bmad-guide` skill in these situations:**
1. **Starting ANY task in a BMAD project** - Check phase and workflow first
2. **Before implementing ANY feature** - Verify which workflow to use
3. **When uncertain about approach** - Consult guide before proceeding
4. **Before creating/modifying code manually** - Should you be using a workflow?
5. **When switching phases** - Verify phase progression rules
6. **When user mentions workflows, epics, stories, or phases** - Get guidance
## BMAD Process Rules (NEVER VIOLATE)
1. **ALWAYS use workflows** - Never implement features without BMAD workflows
2. **NEVER skip phases** - Each phase builds on previous (Phase 1 optional)
3. **ALWAYS check project level (0-4)** - Determines which workflows to use
4. **NEVER jump straight to coding** - Use proper workflow (dev-story, super-dev-story)
5. **ALWAYS follow story lifecycle** - create → dev → review → done
6. **ALWAYS complete current phase** - Before moving to next phase
## Before ANY Work in BMAD Project
\```
1. Invoke /bmad-guide skill
2. Identify current phase (Analysis/Planning/Solutioning/Implementation)
3. Determine project level (0-4)
4. Verify correct workflow for task
5. Check prerequisites are complete
6. Proceed with workflow (not manual implementation)
\```
## Red Flags That Mean You're Off Track
If you catch yourself:
- Writing code without using dev-story/super-dev-story
- Creating PRD/architecture manually instead of using workflows
- Skipping phases or prerequisites
- Not sure which workflow to use
- About to implement without a story
**→ STOP. Invoke `/bmad-guide` immediately and course correct.**
```
## Installation Location
The skill is installed to:
- **User level**: `~/.claude/skills/bmad-guide.md` (available across all projects)
This is intentional so the skill is available in all BMAD projects without needing per-project installation.
## Updating the Skill
When you run `npx bmad-method install` again (to update BMAD), the skill will be updated automatically if a newer version exists.
## Manual Installation
If you want to install the skill without running the full BMAD installer:
```bash
# Create skills directory
mkdir -p ~/.claude/skills
# Copy skill file
cp resources/skills/bmad-guide.md ~/.claude/skills/
```
## Success Metrics
You'll know the skill is working when Claude:
- ✅ Checks phase before starting work
- ✅ Uses workflows instead of manual implementation
- ✅ Follows proper story lifecycle
- ✅ Self-corrects when going off track
- ✅ References the skill when uncertain
- ✅ Asks for clarification about phase/workflow
- ✅ Completes prerequisites before advancing phases
---
**This skill is a core part of the BMAD Method and helps ensure Claude follows the methodology correctly.**

View File

@ -0,0 +1,428 @@
# BMAD Method Guide - Process Navigation & Workflow Selection
You are working within the **BMAD Method (BMM)** - a 4-phase AI-powered agile development framework. This guide helps you stay on track and use the correct workflows.
## 🚨 CRITICAL RULES
1. **NEVER skip phases** - Each phase builds on the previous (except Phase 1 which is optional)
2. **ALWAYS check project level** - This determines which workflows to use
3. **ALWAYS use workflows** - Don't implement features manually without BMAD workflows
4. **ALWAYS consult workflow docs** - Located in `src/modules/bmm/workflows/`
5. **STAY IN PHASE** - Complete current phase before moving to next
---
## 📍 WHERE AM I? - Quick Phase Check
### Phase 0: Brownfield Documentation
**You're here if:** Working with existing codebase that needs documentation
**Workflow:** `document-project`
**Next:** Phase 2 (skip Phase 1)
### Phase 1: Analysis (Optional)
**You're here if:** Need strategic exploration, research, or vision definition
**Workflows:**
- `brainstorm-project` - Explore multiple solution approaches
- `research` - Market/technical/competitive research
- `create-product-brief` - Define product vision
**Next:** Phase 2 (Planning)
### Phase 2: Planning (REQUIRED)
**You're here if:** Need to define requirements and scope
**Workflows:**
- **Level 0-1:** `bmad-quick-flow/create-tech-spec` (quick technical spec)
- **Level 2+:** `2-plan-workflows/prd` (full Product Requirements Document)
- **Optional:** `create-ux-design` (UX/UI specification - after PRD)
**Next:**
- Level 0-1: Skip to Phase 4 (Implementation)
- Level 2: Phase 4 if simple, Phase 3 if needs architecture
- Level 3-4: Phase 3 (Solutioning)
### Phase 3: Solutioning (Conditional)
**You're here if:** Need architectural decisions and epic breakdown
**Workflows:**
1. `create-architecture` - Make architectural decisions
2. `create-epics-and-stories` - Break requirements into implementable stories
3. `check-implementation-readiness` - Gate check validation
**Next:** Phase 4 (Implementation)
### Phase 4: Implementation (REQUIRED)
**You're here if:** Ready to implement stories
**Core Loop:**
1. `sprint-planning` (once) - Initialize sprint tracking
2. **For each story:**
- `create-story` - Create individual story
- `dev-story` or `super-dev-story` - Implement with tests
- `code-review` - Quality review
- Mark story as DONE
3. `retrospective` (per epic) - Review and extract lessons
4. `push-all` - Stage & push all changes when ready
**Advanced/Automated:**
- `autonomous-epic` - Fully automated epic completion
- `super-dev-story` - Enhanced dev with pre/post validation
- `gap-analysis` - Validate story against codebase
- `create-story-with-gap-analysis` - Combined create + validate
---
## 🎯 PROJECT LEVEL DETECTION
**ALWAYS determine project level first** to route to correct planning track:
| Level | Name | Stories | Detection Keywords | Planning Track |
|-------|------|---------|-------------------|----------------|
| **0** | Atomic | 1 | "bug", "fix", "typo", "update", "quick" | Tech spec only |
| **1** | Small | 1-10 | "small feature", "add button", "simple" | Tech spec only |
| **2** | Medium | 5-15 | "feature", "module", "component" | PRD (skip arch) |
| **3** | Complex | 12-40 | "system", "integration", "api", "architecture" | PRD + Architecture |
| **4** | Enterprise | 40+ | "platform", "enterprise", "multiple products" | PRD + Architecture |
**How to detect:**
```bash
# Use workflow-init to automatically detect level
/workflow-init
```
---
## 🔍 WORKFLOW DECISION TREE
```
┌─ START: What do you need to do?
├─ "I need to explore solutions" → Phase 1: brainstorm-project
├─ "I need requirements"
│ ├─ Level 0-1? → Phase 2: bmad-quick-flow/create-tech-spec
│ └─ Level 2+? → Phase 2: 2-plan-workflows/prd
├─ "I have PRD, need architecture" → Phase 3: create-architecture
├─ "I have PRD, need stories" → Phase 3: create-epics-and-stories
├─ "I'm ready to implement"
│ ├─ First time? → sprint-planning (initialize)
│ ├─ Need new story? → create-story
│ ├─ Implement story? → dev-story or super-dev-story
│ ├─ Automate epic? → autonomous-epic
│ └─ Review story? → code-review
└─ "Epic is done" → retrospective
```
---
## ⚠️ COMMON MISTAKES TO AVOID
### ❌ DON'T:
1. **Skip Phase 2** - Requirements are ALWAYS required
2. **Jump to coding** without workflows
3. **Create architecture for Level 0-2** projects (overkill)
4. **Skip sprint-planning** before dev-story
5. **Mix phases** - finish current phase first
6. **Implement manually** - use dev-story/super-dev-story
7. **Forget retrospectives** - extract lessons after each epic
### ✅ DO:
1. **Check phase progression** - Am I in the right phase?
2. **Consult workflow docs** - Read the .md files in workflows/
3. **Use workflow-init** - Let BMAD detect level
4. **Follow story lifecycle** - create → dev → review → done
5. **One story at a time** - Complete before moving to next
6. **Use autonomous-epic** - For efficient bulk processing
7. **Validate with gap-analysis** - Before and after development
---
## 📚 HOW TO LOOK UP WORKFLOW INFO
### Method 1: Read Workflow Documentation
```bash
# Read workflow guide for current phase
cat src/modules/bmm/workflows/README.md
cat src/modules/bmm/docs/workflows-{phase}.md
# Example: Planning phase
cat src/modules/bmm/docs/workflows-planning.md
# Example: Implementation phase
cat src/modules/bmm/docs/workflows-implementation.md
```
### Method 2: Read Specific Workflow
```bash
# Read workflow details
cat src/modules/bmm/workflows/{phase}/{workflow-name}/README.md
# Example: PRD workflow
cat src/modules/bmm/workflows/2-plan-workflows/prd/README.md
# Example: Dev story workflow
cat src/modules/bmm/workflows/4-implementation/dev-story/README.md
```
### Method 3: Check Workflow Configuration
```bash
# See workflow config
cat src/modules/bmm/workflows/{phase}/{workflow-name}/workflow.yaml
```
### Method 4: Use Explore Agent
```
Use Task tool with subagent_type=Explore to search for workflow info:
"Find documentation about {workflow-name} in BMAD workflows"
```
---
## 🎬 TYPICAL PROJECT FLOWS
### Small Change (Level 0-1)
```
tech-spec → dev-story → code-review → done
```
### Medium Feature (Level 2)
```
prd → [optional: create-ux-design] → create-epics-and-stories
→ sprint-planning → (create-story → dev-story → code-review)* → retrospective
```
### Complex System (Level 3-4)
```
[optional: brainstorm/research] → prd → [optional: create-ux-design]
→ create-architecture → create-epics-and-stories → check-implementation-readiness
→ sprint-planning → (create-story → super-dev-story)* → retrospective
```
### Automated Epic Processing (Any Level)
```
[Phase 2-3 complete] → sprint-planning → autonomous-epic {epic_number}
→ (auto: create-story → super-dev-story → commit)* → retrospective
```
---
## 🚀 WORKFLOW EXECUTION PATTERNS
### Pattern 1: Step-File Workflows (PRD, Architecture, Epics)
- **Sequential execution** - Can't skip steps
- **One step at a time** - Load step file, execute, move to next
- **User approval gates** - Wait for approval between steps
- **Append-only building** - Never overwrite previous sections
- **State tracking** - Uses frontmatter `stepsCompleted` array
**Example:**
```
step-01-init.md → step-02-discovery.md → step-03-requirements.md → ...
```
### Pattern 2: YAML Configuration Workflows (Dev, Review, Sprint)
- **Full config in workflow.yaml**
- **Auto-discovery** - Finds story files automatically
- **System execution** - Agent-driven, not user-interactive
- **Config inheritance** - Reads from project config
---
## 🔧 ADVANCED FEATURES
### Autonomous Epic Processing
**When to use:** Want to automate entire epic completion
```bash
# After sprint-planning, automate all stories in epic
/autonomous-epic 2
```
**What it does:**
- Creates story → Develops with super-dev-story → Commits → Repeat
- Includes pre/post gap analysis
- Includes code review
- Creates git commit per story
- Continues on failures (configurable)
### Super-Dev-Story
**When to use:** Want enhanced development with validation
```bash
# Instead of dev-story, use super-dev-story
/super-dev-story
```
**What it does:**
- Pre-dev gap analysis (validate story before coding)
- Implement with tests
- Post-dev gap analysis (validate implementation)
- Built-in code review
- Fail-on-critical-issues (auto-fix loops)
### Gap Analysis
**When to use:** Validate story against actual codebase
```bash
# Standalone validation
/gap-analysis
```
**What it checks:**
- Missing infrastructure before implementation
- Implementation matches story requirements
- All tasks from story are actually completed
---
## 📋 WORKFLOW STATUS & TRACKING
### Check Current Status
```bash
# View sprint status
cat {project-root}/_bmad/bmm/implementation/sprint-status.yaml
# View workflow status
cat {project-root}/_bmad/bmm/workflow-status.yaml
```
### Story States
- `TODO` - Not started
- `IN PROGRESS` - Currently being developed
- `READY FOR REVIEW` - Implemented, needs review
- `DONE` - Complete and reviewed
---
## 🎯 WHEN TO USE EACH WORKFLOW - QUICK REFERENCE
| I need to... | Use this workflow | Phase |
|-------------|-------------------|-------|
| Explore solutions | brainstorm-project | 1 |
| Research market/tech | research | 1 |
| Define product vision | create-product-brief | 1 |
| Write requirements (quick) | bmad-quick-flow/create-tech-spec | 2 |
| Write requirements (full) | 2-plan-workflows/prd | 2 |
| Design UX/UI | create-ux-design | 2 |
| Make architecture decisions | create-architecture | 3 |
| Break into stories | create-epics-and-stories | 3 |
| Validate readiness | check-implementation-readiness | 3 |
| Start sprint | sprint-planning | 4 |
| Create individual story | create-story | 4 |
| Implement story | dev-story or super-dev-story | 4 |
| Automate epic | autonomous-epic | 4 |
| Validate against code | gap-analysis | 4 |
| Review quality | code-review | 4 |
| Review epic completion | retrospective | 4 |
| Push changes | push-all | 4 |
---
## 💡 TROUBLESHOOTING
### "I'm not sure which phase I'm in"
→ Check for existing artifacts:
- PRD exists? You're past Phase 2
- Architecture exists? You're past Phase 3
- Sprint-status.yaml exists? You're in Phase 4
- No artifacts? Start with Phase 2 (or Phase 1 if exploring)
### "I'm not sure which workflow to use"
→ Use this guide's decision tree or consult phase-specific docs:
- `workflows-planning.md` - Phase 2 workflows
- `workflows-solutioning.md` - Phase 3 workflows
- `workflows-implementation.md` - Phase 4 workflows
### "Workflow isn't working"
→ Check these:
1. Is config.yaml set up? (project-root/_bmad/bmm/config.yaml)
2. Are you in the right phase?
3. Did you complete prerequisite workflows?
4. Check workflow README for requirements
### "I want to automate everything"
→ Use autonomous-epic after completing Phase 2-3:
```bash
/sprint-planning # Initialize once
/autonomous-epic 1 # Automate epic 1
/autonomous-epic 2 # Automate epic 2
```
---
## 🎓 KEY PRINCIPLES
1. **Scale-Adaptive** - Same system works for 1 story or 100 stories
2. **Phase-Gated** - Complete phases in order (except Phase 1 optional)
3. **Story-Centric** - Everything revolves around implementing stories
4. **Validation-First** - Gap analysis prevents mismatched implementations
5. **Automation-Friendly** - Can automate entire epics with autonomous workflows
6. **Quality-Built-In** - Code review and retrospectives are mandatory
7. **Documentation-Driven** - Every phase produces artifacts
---
## 📖 ADDITIONAL RESOURCES
### Core Documentation
- `src/modules/bmm/docs/workflows-analysis.md` - Phase 1 guidance
- `src/modules/bmm/docs/workflows-planning.md` - Phase 2 guidance
- `src/modules/bmm/docs/workflows-solutioning.md` - Phase 3 guidance
- `src/modules/bmm/docs/workflows-implementation.md` - Phase 4 guidance
- `src/modules/bmm/docs/scale-adaptive-system.md` - Level detection
- `src/modules/bmm/docs/brownfield-guide.md` - Existing codebases
### Specialized Guides
- `src/modules/bmm/docs/test-architecture.md` - TestArch workflows
- `src/modules/bmm/docs/agents-guide.md` - All 12 specialized agents
---
## ✅ CHECKLIST: Am I Following BMAD?
Before proceeding, check:
- [ ] I know what phase I'm in
- [ ] I know what project level this is (0-4)
- [ ] I've completed prerequisite phases
- [ ] I'm using a workflow (not coding manually)
- [ ] I've read the workflow documentation
- [ ] I'm following the story lifecycle (if Phase 4)
- [ ] I'm validating with gap analysis (if Phase 4)
- [ ] I'm doing code reviews (if Phase 4)
- [ ] I'm completing retrospectives (if finishing epic)
**If any are unchecked, STOP and fix it before proceeding.**
---
## 🚨 EMERGENCY: I Went Off Track
1. **Identify where you are:**
- Check existing artifacts (PRD? Architecture? Sprint status?)
- Determine current phase
2. **Determine where you should be:**
- What was the original request?
- What phase should that be in?
3. **Course correct:**
- Use `correct-course` workflow if in Phase 4
- Otherwise, restart the appropriate workflow for current phase
4. **Prevent future derailment:**
- Consult this guide before starting work
- Read workflow docs before executing
- Use TodoWrite to track workflow steps
---
## 🎯 REMEMBER
**BMAD is a methodology, not a suggestion.**
When in doubt:
1. Check this guide
2. Read workflow documentation
3. Ask the user for clarification
4. Use explore agent to find info
**DO NOT improvise. Follow the process.**

View File

@ -151,6 +151,9 @@ class ClaudeCodeSetup extends BaseIdeSetup {
// Skip CLAUDE.md creation - let user manage their own CLAUDE.md file
// await this.createClaudeConfig(projectDir, modules);
// Install BMAD Guide skill to user's Claude skills directory
await this.installBmadGuideSkill();
// Generate workflow commands from manifest (if it exists)
const workflowGen = new WorkflowCommandGenerator(this.bmadFolderName);
const { artifacts: workflowArtifacts } = await workflowGen.collectWorkflowArtifacts(bmadDir);
@ -194,6 +197,39 @@ class ClaudeCodeSetup extends BaseIdeSetup {
// Method removed - CLAUDE.md file management left to user
/**
* Install BMAD Guide skill to user's Claude skills directory
* This skill helps Claude stay on track with BMAD methodology
*/
async installBmadGuideSkill() {
const os = require('node:os');
// Source skill file in BMAD repository
const sourcePath = getSourcePath('resources', 'skills', 'bmad-guide.md');
// Target directory in user's home
const targetDir = path.join(os.homedir(), '.claude', 'skills');
const targetPath = path.join(targetDir, 'bmad-guide.md');
try {
// Check if source file exists
if (!(await fs.pathExists(sourcePath))) {
console.log(chalk.dim(' Skipping BMAD Guide skill (source not found)'));
return;
}
// Ensure skills directory exists
await this.ensureDir(targetDir);
// Copy skill file
await fs.copyFile(sourcePath, targetPath);
console.log(chalk.green(' ✓ Installed BMAD Guide skill to ~/.claude/skills/'));
console.log(chalk.dim(' Access with: /bmad-guide'));
} catch (error) {
console.log(chalk.yellow(` ⚠ Warning: Could not install BMAD Guide skill: ${error.message}`));
}
}
/**
* Read and process file content
*/